The Model Index
Up to four models

Gemini 3.5 Flash vs Gemini 3 Flash

On the index (GPT-4 = 100), Gemini 3.5 Flash scores 162 and Gemini 3 Flash 154. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.

Next to Gemini 3.5 Flash
Presets
Gemini 3.5 Flash
162146–177
WeightsClosed
ReleasedMay 2026
Price in / out$1.5 / $9.0
Context1.05M
Core tests taken7 of 9
Gemini 3 Flash
154139–170
WeightsClosed
ReleasedDec 2025
Price in / out$0.50 / $3.0
Context1.05M
Core tests taken5 of 9

Gemini 3.5 Flash leads Gemini 3 Flash by 7 points, ahead on 5 of 5 shared core tests, at 3.0× the price.

Test by testthird-party results

Gemini 3.5 FlashGemini 3 Flash
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Epoch AI)GPQA Diamond (Epoch AI)GPQA Diamond (Vals AI)GPQA Diamond (Vals AI)MMLU ProMMLU ProMMMU ProMMMU ProGeoBenchGeoBenchSimpleBenchSimpleBenchSimpleQA VerifiedSimpleQA VerifiedMATHSFrontierMath (tiers 1-3)FrontierMath (tiers 1-3)AIME-style maths (OTIS mock)AIME-style maths (OTIS mock)Chess puzzlesChess puzzlesFrontierMath (2025 set)FrontierMath (2025 set)FrontierMath tier 4 (2025 set)FrontierMath tier 4 (2025 set)FrontierMath (tier 4)FrontierMath (tier 4)CODINGSWE-bench VerifiedSWE-bench VerifiedCode MigrationCode MigrationLiveCodeBenchLiveCodeBenchProgramBenchProgramBenchSWE-benchSWE-benchVibe Code Bench v1.1Vibe Code Bench v1.1WeirdMLWeirdMLAGENTSDTBenchDTBenchLMCALMCAMystery game puzzlesMystery game puzzlesPROFESSIONAL WORKFinance Agent (v2)Finance Agent (v2)CorpFin v2CorpFin v2EMBEMBLegalBenchLegalBenchLegal Research BenchLegal Research BenchMedCodeMedCodeMedScribeMedScribeMortgageTaxMortgageTaxTax Agent BenchTax Agent BenchTaxEval v2TaxEval v2NOVEL PROBLEMSARC-AGI-2ARC-AGI-2ARC-AGI-1ARC-AGI-1
Every result as a table
TestGemini 3.5 FlashGemini 3 Flash
Reasoning & knowledge
GPQA Diamond (Epoch AI)92.8%89.4%
GPQA Diamond (Vals AI)92.7%87.9%
MMLU Pro89.5%88.6%
MMMU Pro88.3%87.6%
GeoBench85.0%88.0%
SimpleBench76.7%61.1%
SimpleQA Verified66.2%66.8%
Maths
FrontierMath (tiers 1-3)62.8%51.2%
AIME-style maths (OTIS mock)95.6%95.6%
Chess puzzles45.0%40.0%
FrontierMath (2025 set)39.0%35.6%
FrontierMath tier 4 (2025 set)14.6%4.2%
FrontierMath (tier 4)26.8%17.1%
Coding
SWE-bench Verified79.3%75.4%
Code Migration26.8%6.4%
LiveCodeBench87.6%85.6%
ProgramBench0.0%0.0%
SWE-bench78.8%75.0%
Vibe Code Bench v1.148.7%20.2%
WeirdML62.6%61.6%
Agents
DTBench94.7%89.1%
LMCA47.1%43.1%
Mystery game puzzles32.0%26.0%
Professional work
Finance Agent (v2)57.9%42.5%
CorpFin v264.7%66.4%
EMB63.5%26.0%
LegalBench83.6%86.9%
Legal Research Bench30.8%18.3%
MedCode55.8%55.9%
MedScribe76.6%69.9%
MortgageTax68.1%68.7%
Tax Agent Bench21.6%11.8%
TaxEval v274.4%73.9%
Novel problems
ARC-AGI-272.1%33.6%
ARC-AGI-192.5%84.7%

Tests at least two of these models have taken. Bold rows are the core tests.

This page as a chartDownload PNG